Panorama Generation from a Video

Video Panorama is a complete view of an object that is generated by aligning and blending a sequence of images. A video sequence consists of several hundred frames, and as a result, creating a panoramic image from these frames is a challenging task because video may have moving objects and the objects may appear at different depths. Computer vision is a field that includes methods for acquiring, processing, analyzing, and understanding images and, in general, high-dimensional data from the real world in order to produce numerical or symbolic information. One of the sub domain is generating panoramas from video which have applications in aero photography, satellite generated photographs to be stitched for research. Our aim is to design a Panorama Generation System that takes a video as input and captured by a moving camera navigating through a scene and our goal is to generate a panoramic summary of the scene and for this purpose remove moving and foreground objects. The panorama process receives two dimensional video from video camera. The received video includes set of image frames and the frames together form a panorama image. To build panorama image from video, we use the motion model based on color invariant features in order to apply our algorithm in a scene where there are illumination, highlights and shadows. After that, key frames/regions are selected for the following image fusion process. At last, the input frames are transformed to the same coordinate to fuse a panorama.